Linear Solvation Energy Relationships For The Partition Coefficients Of Organic Contaminants Between Semipermeable Membrane Devices And Environmental Media

Semi-permeable membrane devices (SPMD) are employed as a passive sampler in order to determine environmental levels of organic contaminants. The sampling material is the triolein contained in the tube. The equilibrium partition coefficients between the sampling material and the environmental media (SPMD/air partition coefficient KSA and SPMD/water partition coefficient KSW) are required in order to determine the sampling rates and estimate the concentrations of pollutants.The further application of SPMD is hindered partly due to the lack of available KSPMD values. In order to predict the values of the KSPMD, LSER (linear solvation energy relationship) theory was adopted to establish predictive models. The LSER molecular descriptors were collected and calculated to develop the LSER models in order to predict the KSPMD.(1) The LSER molecular descriptors were calculated and collected for some common semi-volatile organic contaminants, including PAHs (polycyclic aromatic hydrocarbons), PCBs (polychlorinated biphenyls), chlorinated phenols, organochlorine pesticides, pyrethroids, organophosphorus and organosulfur pesticides, and MLR (multivariate linear regression) was employed for model development. Statistical parameters (R2 adj, SE, F, p and VIF) showed the developed LSER model had good fitting ability, robustness and predictivity. Moreover, the Eucildean distance-based approach was adopted to study the AD (application domain) of the model. The factors governing the partition were the diffusion effect (negatively correlated with KSA) and molecular dipolarity/polarizability (positively correlated with KSA), between which, the KSA was more strongly affected by the molecular dipolarity/polarizability. The superiority of the established LSER models, as compared with those built by others, lied in the interpretability of the mechanisms and the availability of molecular parameters.(2) The LSER molecular descriptors were calculated and collected for some common hydrophobic organic contaminants, including (chlorinated) benzenes, PAHs, PCBs and pesticides, and MLR was employed for model development. Statiscal parameters (R2 adj, SE, F, p and VIF) showed the developed LSER model had good fitting ability, robustness and predictivity. Moreover, the Eucildean distance-based approach was adopted to study the AD of the model. The factors governing the partition were the diffusion effect (positively correlated with Ksw), H-bonding (negtively correlated with KSW) and cavity formation (positively correlated with KSW).